rdzidlic writes

#1: firefox extensions use xml for many purposes and they may cause errors like this:

mozilla/home-changer/content/options.xul:6: parser error : Entity 'hc.text.preferences' not defined
    <setting type="control" title="&hc.text.preferences;" >
                                                        ^
mozilla/home-changer/content/options.xul:7: parser error : Entity 'hc.text.open' not defined
        <button label="&hc.text.open;" oncommand="window.open('chrome://homechan
                                     ^
mozilla/home-changer/content/preferences.xhtml:7: warning: failed to load external entity "chrome://homechanger/locale/preferences.dtd"
%hcPrefDTD;
           ^
Entity: line 1:
 %hcPrefDTD;
            ^
mozilla/home-changer/content/preferences.xhtml:13: parser error : Entity 'hc.title.page' not defined
  <title >&hc.title.page;</title >
                        ^
mozilla/home-changer/content/preferences.xhtml:24: parser error : Entity 'hc.title.header' not defined

#2 OsmAnd xml causes this errors:

android/osmand/resources/obf_creation/rendering_types.xml:03: parser error : attributes construct error
tor" if_value1="東武鉄道;東京地下鉄" if_tag2="route" if_value2="subway"

android/osmand/resources/obf_creation/rendering_types.xml:03: parser error : Couldn't find end of Start Tag entity_convert line 2003
tor" if_value1="東武鉄道;東京地下鉄" if_tag2="route" if_value2="subway"

medoc writes

My best guess would be that these files are not valid xml, but it’s difficult to say without a sample. Actually I sometimes think that it would be best to just index XML as text. For example, the current code does not index attributes at all.

You can change this easily by adding the following to $HOME/.recoll/mimeconf:

[index]
application/xml = internal text/plain
text/xml = internal text/plain

rdzidlic writes

treating xml as text seems like a good idea. Otherwise it is infinite madness, even domain specific tools like gpsbable often fail on kml files.

medoc writes

Yes, I agree. The next version will use the above setup by default for XML.

medoc writes

Index generic XML as text by default. Closes issue #277

→ <<cset 85b157d11b5c > >